file.pdf (328.78 kB)
Download file

Mathematical Method and Proof

Download (328.78 kB)
journal contribution
posted on 01.01.2006, 00:00 by Jeremy AvigadJeremy Avigad
On a traditional view, the primary role of a mathematical proof is to warrant the truth of the resulting theorem. This view fails to explain why it is very often the case that a new proof of a theorem is deemed important. Three case studies from elementary arithmetic show, informally, that there are many criteria by which ordinary proofs are valued. I argue that at least some of these criteria depend on the methods of inference the proofs employ, and that standard models of formal deduction are not well-equipped to support such evaluations. I discuss a model of proof that is used in the automated deduction community, and show that this model does better in that respect.


Publisher Statement

The original publication is available at



Usage metrics